Solder Mask

[PixieDust]

Hi guys, is solder mask doable at home? All the videos on Youtube that I've seen produce unacceptable results.
I think the most impressive results are from laminating dry film solder mask exposing to UV and etching. But I'm not sold on the idea of this type of solder mask. I don't know how well or poorly it performs in the long run. Does it de-laminate over time? And then how good is it standing up to the heat of soldering?

[jmelson]

I have some scraps of PC board material with DuPont Riston resist that were laminated probably 20 years ago, and the resist is still adhering as well as it did when first exposed and developed.  Now, some of it might break down right around where you have soldered components on, but in general, it is really tough and stable.  They do sell a similar material to be used as a solder mask.  The only reason it is rarely seen is it is pretty expensive, so production shops use cheaper stuff.

Jon

[sleemanj]

The short answer is: No.  Don't bother.  It is a total PITA.  When you get to the point of needing a soldermask, you are at the point where spending $20 for to get your PCBs made in China and in your hands next week is worth it.

UV ink is messy, awkward, finicky and difficult at home
Dry Film Mask (don't confuse with resist) is not that readily available and available in green only
Vitrea 160 with toner masking is semi-workable but still a tricky process that needs considerable practice and experimenting

And after all that you still just have a PCB with a usually DIY looking soldermask and no silkscreen.  You might just as well paint the PCB manually with nail varnish since that will be about as protective.


 

[Kean]

You might just as well paint the PCB manually with nail varnish since that will be about as protective.

Agreed.  Or just use some Kapton tape.

[Mark]

Yes it is possible at home.  I use Dynamask 5000 series Dry Film Solder Mask in green.  It takes about 60 seconds for exposure using a proper UV unit (design for PCB use).  After developing, it can be hardened further by exposure to UV for 10-30 minutes or so, or pop it in the reflow oven at 100C for a while. 

I've also had good results with the UV curable paint on eBay.  Buff the PCB with a wire brush or scotchbrite pad first to help with adhesion, then apply a small drop on the PCB.  Place a shiny acetate sheet over the top and spread the drop out until there is a thin and even layer of paint.  Then align the artwork (keeping the original shiny acetate sheet in place) and expose for 45 seconds.  Gently wipe with IPA and a soft cloth to remove unexposed paint, then harden by exposing for another few minutes. 
Use disposable gloves, it's not all that messy. 

[Mark]

This was my first attempt with the paint, blue on one side and yellow on the other.  That board was double sided CNC milled.   

The green board is dynamask (not properly hardened before scrubbing with flux cleaner brush).  That board was photo-imaged then etched. 

[KL27x]

And after all that you still just have a PCB with a usually DIY looking soldermask and no silkscreen.  You might just as well paint the PCB manually with nail varnish since that will be about as protective.

On an SMD board, in particular, the soldermask is useful for keeping the solder on the pads, rather than wicking anywhere it wants.

I would not call the results impressive, but I have learned how to do a very easy "solder mask" on a home made board by growing an oxide layer on the copper.

What I do is take the finished board and shine it up. I wet scrub with stainless steel wool and then pat it dry with paper towel. Then I take the board over to my microscope and draw over the pads with sharpie. The clearance around the pads helps to make this relatively quick and painless. Then I dip the board in my etchant for a few seconds. Rinse thoroughly in clean water. Then hang it up, wet. In some number of hours, something between 0 and 6 (dunno, because I never sat there and watched), the board turns black.

Wipe the board with alcohol or acetone, and you get shiny copper pads on a black board. I'm way more careful of how much/where I apply the liquid flux when I solder. There is a heat affected zone around where you apply your iron to fluxed board, where the oxide layer starts to dissolve. But with a minimal amount of care, the solder stays 100% on the pads. Even where I forget to apply a thermal pad on a plane, this dot of clean copper can take a good hot reflow without spilling out of bounds. Even a mil or two line of oxide that is incidentally made when I sharpie adjacent pads holds up and keeps the solder to its respective pad. Population is also much easier, because the pads really pop out at you. Esthetically, I find the result very pleasing, esp after cleaning and a layer of conformal coat. But it's not for everyone. It looks a bit like the shell of an old crusty blue lobster.

I have some pics:
[Imgur](

)
After soldering and handling, some of the black will be lost. Here's the final, after cleaning but before the conformal coat. You can see how the solder stays on the pads, compared to the unmasked board.
[Imgur](https://i.imgur.com/BNBL4uJ.jpg?1)
After a conformal coat:
[Imgur](https://i.imgur.com/7lsCju8.jpg?1)

Note this board was initially etched over 7 years ago. There was some pitting and some rough traces, although it might be hard to see with the oxide layer. My transfer and etch is much cleaner these days.

Anywhere you forgot to cover the pad or where you want to add a pad to a ground plane, you can scrape away a spot on the oxide layer. In fact, I first started playing with this idea on blank copper clad which I use for adding a component or two to existing board by mechanically cutting pads/traces. Then I'd reveal where I want the pads to be by removing the oxide only at those spots.

[PixieDust]

On an SMD board, in particular, the soldermask is useful for keeping the solder on the pads, rather than wicking anywhere it wants.

This is probably the principal reason for looking into this. Protecting the board is nice for a piece of mind too.

[KL27x]

^Sounds like you are not happy with the many commercial solutions, though? For esthetics and durability reasons? "Unacceptable results" and concerns over "long term durability?"

They appear to all work pretty well, to me. The major issue is the work involved!

Give my method a try, sometime. I'm curious how it works for someone who does oven reflow. I have a reflow oven, but I have no way to make stencils. I also have no desire to do it, anymore, because I am not good at pasting a stencil or using paste applicators.

[Mark]

Give my method a try, sometime. I'm curious how it works for someone who does oven reflow. I have a reflow oven, but I have no way to make stencils. I also have no desire to do it, anymore, because I am not good at pasting a stencil or using paste applicators.

Stencil making is quite easy, all you need is a Silhouette Portrait or Cameo craft cutter, some mylar sheets gerber2graphtec (python scripts) .  Export the paste file as a gerber and import it into the gerber2graphtec program, then send it to the printer.   

https://github.com/pmonta/gerber2graphtec

The craft cutter chugs away for 10-30 minutes depending on the complexity of the design and the number of rounded corners used. It works better/faster with squares and rectangles than with circles.  Converting circles to octagons is advised if you want to save time. 

VQFN-56 (8mm×8mm) is possible but you may need to remove some of the "hanging chads" using a pair of tweezers. 

[SMTech]

Hmmm I've tried laser cut polyester stencils once and I was not impressed at all. Unlike a metal stencil you can't apply any tension because it deforms and seemed to do so simply from the drag of the squeegee. Mylar might be better as structure is different but the other problem i saw was the apertures were really bad at releasing the paste on smaller apertures.

Fine for nothing that fancy and small boards in one or twos but frankly for anything else I think you'd have to be off your rocker. The idea on here https://www.smtstencil.co.uk/ that people use them 100s or even 1000s of times is incomprehensible to me. Not having a frame is bad enough in my book, were I a hobbyist I would be making something that could act as a frame post haste.

[ar__systems]

What I do is take the finished board and shine it up. I wet scrub with stainless steel wool and then pat it dry with paper towel. Then I take the board over to

I can't imagine why would anybody want to do this, other than for fun. Commercially made board are cheap.

[KL27x]

I can't imagine why would anybody want to do this, other than for fun. Commercially made board are cheap.

*Shrug* It seems to me it's a lot easier than the UV paint/film methods. What part about that sounded as labor intensive as say cleaning up a paste stencil after using it? The wet scrub is already part of my process for removing the toner and revealing clean copper. The only real added step is coloring in the pads with the sharpie and dunking it, again. I suppose at some point in size/complexity, UV masks will be easier.

All you need is a cutting board and a sink. You've got running water, right? It's like doing the dishes except you're done after one dish.

I spent a few long days doing the board layout (it's double-sided with a high density bottom layer). I spent hours figuring out how the low voltage cutout pins on the 1308 work and writing code for the micro for on/off and low voltage cutout and microamp sleep. I spent hours initially sourcing the parts that I didn't have on hand. I spent hours and two evenings assembling the board, mostly in locating and handling the parts. My process for making PCB takes some time and work, but it's highly optimized and a fairly small part of total work and time on this particular board. The added time/work allows me to do the other stuff on my own schedule rather than waiting for a package to show up. The hard work of tweaking the transfer equipment and making the tank/etchant has already been completed as a one time investment. (The dedicated cutting board was part of it. Highly recommended.) If something were to break the camel's back, it wouldn't have been the manufacturing of the PCB or the 60 seconds it took to color in the pads with a Sharpie.

This size of board, I can make probably 6-8 in a go. This was the one leftover from a batch I made 6-7 years ago. I probably kept it bare for reference/debugging while populating the other ones. I was running low on this circuit after giving some away over the years. After making this one and repairing and reworking one of the earlier protos to bring it to speed, I'm good, again.

[ar__systems]

I was referring to the whole process of hand made boards, not just the mask. Yes, you may argue the few hours you spent on the PCB is 1% of the entire product development time, but it is 1% easily avoided. 

[KL27x]

You are asking; I am answering. There's no arguing, here. If you don't like what I'm saying, then stop reading. If you want to convince me to stop making PCB's, I'm all ears.

I order boards, too. They can be very inexpensive, but it's usually at least 15-20.00. Add another 20, if you want 3 day shipping.

To come clean, here, most of my homemade boards are not anywhere as complex as this example. More often, I am using toner transfer to make adaptors, breakout boards, and for testing new circuitry/components with fairly simple single sided boards just to add some support circuitry and pin headers to an IC footprint. For this kind of thing it is rare that any copy beyond the first one has any value to me, whatsoever. So that 15 to 20 bucks is not insane, but for a single breakout board that ISn't going to arrive in two days? Maybe I'm cheapskate, but that's not great to me.

I might use another one in a year or 10 (not likely), but that's only if I can find it. If I make my own, I just have to find a file on my computer. An analogy is how I deal with my component storage. It's not fast and sleek. It's very compact and secure, but it takes time to find things and to get the parts out. No ziplock baggies, no slide out compartments. I just seal the components up with a heat sealer. I have to cut the bag open and dot it closed again with my iron. But the idea is to store more things, securely, in less space. The "ease of use" is secondary. Ordering boards is like opening the package and taking out ten. Then having no way to put the genie back in the bottle. Now the extra boards are my problem. That one use cost 20-30 dollars. I can't count the number of professionally manufactured PCBs I have thrown away. The $2 per way of thinking is nonsense. When I order 10 pcb, I often thrown 8 or 9 of them away without populating them. Hopefully sooner, rather than later. My workspace thanks me. On work related proto, maybe 3 sets of boards ever gets used, depending on how many parties need a copy. This is just for the electrical verification and housing. If there are failure rate/assembly issues to work out and streamline, it is gonna take more than 10 copies and usually requires the professionally populated first run. So the extra proto boards are usually as useful as a second dick.*

Making boards is not for everyone. But I think it' a lot easier than many people think. Toner transfer can be 100% reliable and has good enough resolution for most double-sided non BGA things.

*One of the few things I have repeatedly reached for and use up are little DC boost circuit PCBs. This is also, not coincidentally, a pre-populated "component-on-PCB" that you can buy from component distributors. There are probably a few other good examples, but I don't have any for what I do. Maybe Arduino's and stuff like that. RS232 adapters. I dunno. I mean, if it's a board you are gonna find useful again and again, someone probably makes and sells it cheaper than you can.

[Siwastaja]

I make a board when I need it in 2 hours. Doesn't happen usually now, but when it does happen, it does.

Oh, this is easy to calculate.

If I value my work worth of $300/hr, and I work 12 hours a day or 60 hours a week when in flow, and if we assume that I can't do significant substitute work on the project but need to prototype the electronics ASAP - does happen sometimes - we get:

DIY board: $600 worth of my time and $2 worth of materials -> $602
Chinese $10 board (cheap shipping): 3 weeks: $54010
Chinese $50 board (DHL shipping): 1 week: $18050
Local quick turn-around service board: $500 and 2 days: $7700

"LOL DON'T MAKE YOUR BOARDS THEY ARE SO CHEAP" spammers are the cancer of this forum, year after year. They automatically derail and ruin every topic with a bot-like 100% certainty, and are always stupid enough not to understand long and detailed expert replies they get, so it's always time wasted trying to discuss with them.


On topic: I don't usually add a solder mask; it's at least one hour more, and the quality is never on par with the fabs. I rather spend that one hour Kapton taping the one single spot which is susceptible; often there is no need at all for this. Same goes for through hole plating; I achieved around 99.5 to 99.9% yield, which is order of magnitude too low, and spent 1.5 hours per board in plating activities. In the same time, you solder wires to several hundred vias! In any case, I'm fairly happy with green dry film solder mask material, specifically for this purpose. You just need to expose it longer than similar etch photoresist - otherwise, the surface is damaged during the carbonate developer step. Post exposure is also highly recommended.

[PixieDust]

Yes, I would rather not have this thread derail into a discussion on should one make a board at home or order one online. That's not my intent.

Does anyone have any suggestions as to what thickness soulder mask should be? I'm leaning in the direction of getting a process happening using liquid stuff. Google isn't yielding much results. Maybe I'm looking for too hardcore info on all this in the form of scientific papers, but would be nice to know some actual scientific reasons for optimal solder mask in various applications.

Also, does anyone have any way of measuring solder mask thickness apart from buying a probe apparatus as seen below, which I can only imagine is prohibitively expensive:

https://www.fischer-technology.com/en/united-states/knowledge/application-notes/premium-measurement-instruments-for-semiconductors-and-electronics/controlling-the-thickness-of-solder-resist-in-the-manufacture-of-printed-circuit-boards/

I can presently use calipers but they only have 0.01mm resolution which means that according to google results, people have seen 0.4 to 1.2mil thick solder mask which is a range of 0.01-0.03mm. Seems too inaccurate? No?

The other alternative is to use a DTI, but that has a similar resolution of 0.01mm.

[brabus]

I make a board when I need it in 2 hours. Doesn't happen usually now, but when it does happen, it does.

Oh, this is easy to calculate.

If I value my work worth of $300/hr,
DIY board: $600 worth of my time and $2 worth of materials -> $602
(...)

All the time spent developing the process before refining it to the point of having a good board in two hours should be taken into account as well, though it gets amortized through the number of boards produced.

[Fred27]

I've got some Dynamask 5000 that I'm unlike to use. I'd anyone in the UK wants some (or if anyone outside the UK wants to pay postage) I'm happy to sent out some pieces to give it a try.

[ar__systems]

Yes, I agree it is annoying to throw out boards. I would like to avoid that, if I could. But I don't want a procedure that is now fire-and-forget to turn into a project of itself....

Lol. Dude, your time is so valuable yet you have nothing else to do in the time that you wait for boards to arrive?

[Siwastaja]

Lol. Dude, your time is so valuable yet you have nothing else to do in the time that you wait for boards to arrive?

Not necessarily my time. And this doesn't always happen. I do a DIY PCB maybe once every two years now.

But, sometimes a project is timing-critical. Sometimes a quick prototype is really necessary to move on. Heck, it can be ten engineers waiting, each at $300/hr. It can be 100 engineers and a multi-billion business opportunity waiting. The numbers are completely arbitrary, I could have picked any unimaginable numbers, but I didn't. The point is, you can't always find substitute work in certain style of projects, and your project is stalled, and this is not going to be free - you are waiting to make some tests and measurements, to prove a concept, to show a prototype - all needed to go on early. In that regard, my numbers are very conservative, assuming only one engineer's time, and not assuming much more value to the time than what many freelancers get paid. Note that there is a difference between your wage, what it costs to the company, and what its value is. Especially the last one can be much more.

For this exact reason, large corporations doing design do have their own in-house prototyping fabs! For this reason, professional PCB milling machinery exist!

Many of us have hobbyist background. I started doing PCBs myself, as a hobby, when the PCB fab cost was ten times more than what it is now. Now, the skill and equipment is very good to have.

Sorry for replying to a derailer troll. I'm not replying anymore after pointing this out.

[ar__systems]

Heck, it can be ten engineers waiting, each at $300/hr. It can be 100 engineers and a multi-billion business opportunity waiting. The numbers are completely arbitrary

Your numbers are not arbitrary. They are pulled out of you know where. The project involves 100 engineers @ 300/hr and they all are waiting for a poor fellow to hand made a circuit board. Right, that happens every day.

[KL27x]

But I don't want a procedure that is now fire-and-forget to turn into a project of itself....

The making of a PCB can be more of a process than a project. At least, if you happen to figure out how to do it well. A project implies a challenge that requires decision making and judgment calls and experience, and which can sometimes result in failure.

Yeah, it's not for everyone. If you do most of your electronics work in an office building and leave your work there when you go home, then it could be highly impractical to even handle chemicals at your workplace. This is probably part of the reason some companies pay thousands of dollars for CNC machines for the primary purpose of making prototype pcbs. If you are a salaried/hourly employee, your manager surely wants to see you at your computer 6 hours a day, minus the 2 hours of mandatory department and company meetings and progress reports. In the typical workplace environment, it is probably safer to sit at your computer and discretely play solitaire rather than be seen making a PCB in the breakroom sink. If you don't have a space where you can conveniently store and use the equipment at your own discretion, then it might turn into a risky project.

and they all are waiting for a poor fellow to hand made a circuit board. Right, that happens every day.

I'm not saying this meets this exact example, but one place where something sorta like this might happen is with testing equipment. For a product that is in production, it may need special jigs for testing during production. Say some pogo pin getup in some sort of press or fixture. Say the new guy broke the jig. Then the manager comes over and expertly installs the new one. But he's an idiot. Now the entire line can go down. Workers paid but no work. Contracts/deadlines in jeopardy and/or lost sales. This can cost a lot of money. Any specialized equipment for production will qualify.

Maybe Siwastaja has gotten a phone call. "Hey, remember that thing you made for use in 2012? Yeah, that it. Do you have another? Oh, how fast can you make another? We need it, now. Someone can drive over and pick it up this afternoon."

Some companies you deal with, everything is a rush. It's always "How fast? You know it would really help us out if you can do in [stated time x0.5] Oh, well, in case it's done sooner, it will help us out. And overnight it. AM Express." You wonder sometimes if it's a personality disorder. But sometimes it's maybe an actual emergency.

[ar__systems]

Some companies you deal with, everything is a rush. It's always "How fast? You know it would really help us out if you can do in [stated time x0.5] Oh, well, in case it's done sooner, it will help us out. And overnight it. AM Express." You wonder sometimes if it's a personality disorder. But sometimes it's maybe an actual emergency.

That might happen, although not in my experience. The companies I deal with know the critical pieces of the process and always have at least two of those in operation, AND backups.

Emergencies like that are not true emergencies, rather they are result of managerial screw up. If that should happen to me I most definitely will not drop everything to start hand etching a PCB. I don't want a sketchy hand made PCB to be used in a production, and then risk being responsible for the failure. If they have to wait for a properly made PCB a day or two, then so be it.

[KL27x]

The companies I deal with know the critical pieces of the process and always have at least two of those in operation, AND backups

There are all kinds of people and personalities that make it into all kinds of positions based on various factors which sometimes includes competence. There are also occasional mistakes by otherwise competent people.

Emergencies like that are not true emergencies, rather they are result of managerial screw up. If that should happen to me I most definitely will not drop everything to start hand etching a PCB.

And yet, when a managerial screw up happens, all the guys that showed up every day, worked hard at their job, and did NOT cause the screw up, they can also lose hours or even be laid off. This company's business might also be a significant portion of your own company's income and ultimately your own livelihood. So yeah, in a "managerial screw up" like this, I might be happy to accommodate. 

If you have the capability to do something that will save/make someone else buckets of money, is that not an opportunity? You should like to be in this situation. For the warm fuzzy feeling of helping other people feed their kids. Or for job security and leverage in your own compensation. Take your pick.

But don't take this as a suggestion from me to make your own PCB's. In fact, I suggest you, AR specifically, do not try it. If it makes you unhappy to even think about it, then don't do it.